Cells with tumorigenic potential are abundant in MPNST
Malignant peripheral nerve sheath tumors (MPNSTs) are soft tissue sarcomas, which form large masses within peripheral nerves, causing pain, abnormal sensations and other neurological deficits. MPNSTs can develop spontaneously or in the context of Neurofibromatosis (caused by Nf1 mutations), an often-hereditary syndrome which predisposes patients to cancers as well as benign tumors of the nervous system. About 30% of NF1 patients develop benign plexiform neurofibromas associated with major nerve trunks, which can undergo malignant transformation to MPNSTs. MPNSTs are very aggressive and highly invasive.1 Neurofibromas and MPNSTs develop from Schwann cells. During tumor formation, some Schwann cells degenerate, driving other Schwann cells into cycle. This recruits inflammatory cells into the nerve, ultimately leading to the formation of large tumors.2,3,4
Research over the past several years has shown that in some cancers a small subpopulation of tumor cells is responsible for tumor growth. These so-called cancer stem cells contrast with the majority of tumor cells, which have a limited capacity to divide or to contribute to tumor growth.5,6 The cancer stem cell model suggests fundamentally different strategies for treating cancer. Cancers that follow a stem cell model might be more effectively treated by focusing therapy on the small subpopulations of cancer stem cells, to more effectively eliminate these cells. In contrast, cancers that do not follow the stem cell model (cancers in which many cells are capable of contributing to tumorigenesis) would require therapies that eliminate as many cancer cells as possible. Therefore, a fundamentally important question is whether only rare subpopulations of MPNST cells are capable of contributing to tumor growth or whether many MPNST cells contribute to tumorigenesis.
With the generous support of the Liddy Shriver Sarcoma Initiative, we set out to test whether tumor-initiating cells were frequent or rare in MPNSTs. Since human tumor material was not readily available (given the rarity of MPNSTs and the difficulty associated with sharing live tumor cells intra-institutionally, immediately after surgery), we took advantage of two mouse models of MPNSTs, which carry similar mutations as found in human MPNSTs. MPNSTs often arise in patients that lack one functional copy of the Neurofibromin 1 (Nf1) gene. The cells that give rise to MPNSTs often loose the other copy in the Schwann cells that give rise to the tumors. Subsequently other genetic changes occur in the tumor cells, leading to the loss of additional genes that negatively regulate cell growth. This leads to the formation of MPNSTs. 25-40% of mice that are heterozygous for Nf1 andp53, or that are heterozygous for Nf1 and lack Ink4a/Arf develop MPNSTs, predominantly on their legs, backs and shoulders.7,8,9 These tumors contain large numbers of proliferating cells (Fig 1).
We developed systems to determine the frequency of tumorigenic cells in mouse MPNSTs of both genotypes (Figure 2). We dissociated the tumor using enzymes to break down extracellular materials and release the cells. We then removed non-cancerous supporting cells such as blood and endothelial cells from the single cell suspensions by flow cytometry, taking advantage of the fact that these cells express different surface molecules than the tumor cells. To determine the frequency of tumor-initiating cells, we cultured single tumor cells in liquid culture medium supplemented with nutrients. After 14 days we found that many of the single tumor cells had divided multiple times and given rise to spheroid colonies, containing hundreds to thousands of cells. We transplanted these spheres under the kidney capsule of mice and monitored the mice for tumor formation. To confirm that culturing the cells in an incubator had not changed the tumor cells, we also injected small numbers of freshly isolated tumor cells into mice without prior culture and observed the frequency of cells that were capable of forming a tumor.
Using the assays we developed, we found that many MPNST cells are capable of forming tumors when transplanted into recipient mice. Indeed, even when we transplanted single MPNST cells into normal mice, 15-20% of the cells from tumors of each genotype were able to form tumors. This demonstrates that many MPNST cells are capable of forming tumors, and that MPNSTs do not follow a cancer stem cell model characterized by the presence of only rare cells with tumorigenic capacity. This suggests that it will not be possible to improve the treatment of MPNSTs by targeting rare subpopulations of MPNST cells, as has been proposed in the context of the cancer stem cell model. Rather, many MPNST cells have the potential to contribute to tumor growth and disease progression meaning that therapies will have to eliminate most or all cells. We are currently preparing a manuscript describing our results for publication in a scientific journal.
Having determined that mouse MPNSTs contain a large number of tumorigenic cells, we would like to find out more about the molecular mechanisms that the tumor cells use to proliferate in an uncontrolled manner. Since MPNSTs arise from Schwann cells, which are usually quiescent, we would like to determine which signaling pathways are switched on in tumor cells to allow for extensive proliferation and whether they resemble the signaling pathways used by neural crest stem cells. To address this issue we will compare the gene expression profiles of mouse MPNST cells to normal neural crest stem cells and differentiated Schwann cells. We then plan to perform functional studies on particular pathways that we think might regulate the proliferation of MPNST cells.
Summary and Conclusion
We found that mouse MPNSTs contain many cells with the potential to proliferate extensively and form tumors. Our results suggest that MPNSTs are different from cancers, such as leukemias, that are driven by rare cancer stem cells. With our future studies, we hope to elucidate molecular mechanisms necessary for MPNST cell survival and proliferation. Our studies might lead to the identification of new therapeutic targets.
By Johanna Buchstaller, PhD
and Sean J. Morrison, PhD
Howard Hughes Medical Institute
Center for Stem Cell Biology, Life Sciences Institute
Department of Internal Medicine
University of Michigan in Ann Arbor, Michigan
1. Ferner RE, Gutmann DH. (2002). International consensus statement on malignant peripheral nerve sheath tumors in neurofibromatosis. Cancer Res 62:1573-1577.
2. Joseph NM, Mosher JT, Buchstaller J, Snider P, McKeever PE, Lim M, Conway SJ, Parada LF, Zhu Y, Morrison SJ. (2008). The loss of Nf1 transiently promotes self-renewal but not tumorigenesis by neural crest stem cells. Cancer Cell 13:129-140.
3. Wu J, Williams JP, Rizvi TA, Kordich JJ, Witte D, Meijer D, Stemmer-Rachamimov AO, Cancelas JA, Ratner N. (2008). Plexiform and dermal neurofibromas and pigmentation are caused by nf1 loss in desert hedgehog-expressing cells. Cancer Cell 13:105-116.
4. Zheng H, Chang L, Patel N, Yang J, Lowe L, Burns DK, Zhu Y. (2008). Induction of abnormal proliferation by nonmyelinating schwann cells triggers neurofibroma formation. Cancer Cell 13:117-128.
5. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB. (2004). Identification of human brain tumour initiating cells. Nature 432:396-401.
6. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. (2003). Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci USA 100:3983-3988.
7. Quintana E, Mark Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morrison SJ. (2008). NOD/SCID mice substantially underestimate the frequency of melanoma-initiating cells. Nature in press.
8. Cichowski K, Shih TS, Schmitt E, Santiago S, Reilly K, McLaughlin ME, Bronson RT, Jacks T. (1999). Mouse models of tumor development in neurofibromatosis type 1. Science 286:2172-2176.
9. Vogel KS, Klesse LJ, Velasco-Miguel S, Meyers K, Rushing EJ, Parada LF. (1999). Mouse tumor model for neurofibromatosis type 1. Science 286:2176-2179.
V7N5 ESUN Copyright © 2010 Liddy Shriver Sarcoma Initiative.
Do many or few cells within malignant peripheral nerve sheath tumors have the potential to contribute to disease progression?
Neurofibromatosis Type 1 is a hereditary tumor syndrome with an incidence of approximately 1 in every 4000 people.1 It is caused by mutations that reduce the function of the Neurofibromin (Nf1) gene on chromosome 17. Most patients inherit these mutations from their families, but some patients acquire spontaneous new mutations. The Nf1 gene encodes a protein that reduces cellular signals, through the Ras pathway, that promote proliferation. When mutated, Neurofibromin’s ability to regulate cellular proliferation is reduced or eliminated. About 30% of patients with Neurofibromatosis Type 1 develop plexiform neurofibromas,2 benign tumors that arise from Schwann cells in nerve fibers and that can grow quite large.3,4 Additional mutations allow these plexiform neurofibromas to progress to malignant peripheral nerve sheath tumors (MPNSTs) in some patients. These soft tissue sarcomas are very aggressive and therapy resistant.
During embryonic development, Schwann cells arise from neural crest stem cells that migrate into developing nerves (Figure 1). Signals from developing axons regulate the migration, proliferation and differentiation of neural crest stem cells into Schwann cells, then induce myelination shortly before birth. Myelination is the process by which Schwann cell processes ensheath axons to insulate them, facilitating the transmission of nerve impulses. Schwann cells usually remain quiescent in nerves; however, after a nerve injury, Schwann cells can dedifferentiate under the influence of axonal signals and resume proliferation. We have recently shown that plexiform neurofibromas and MPNSTs in mouse models bearing Nf1 mutations arise from Schwann cells.4 Independent studies from other laboratories have confirmed this conclusion and suggest that abnormally differentiated non-myelinating Schwann cells degenerate postnatally, inducing inflammation and leading to the development of neurofibromas.3
The goal of our current research is to identify which cells within MPNSTs drive tumor growth. In recent years, a number of cancers have been shown to follow a "cancer stem cell model" in which not all cancer cells are created equal, even within the same tumor (Figure 2).5,6 According to this model, tumors can be thought of in much the same terms as normal tissues. Just as normal tissues contain small subpopulations of stem cells that drive tissue growth by giving rise to phenotypically diverse populations of mature cells, tumors sometimes contain small subpopulations of cancer stem cells that drive tumor growth by giving rise to phenotypically diverse populations of non-tumorigenic cancer cells. The fundamental idea is that just as normal stem cells can differentiate into mature cells with limited proliferative capacity, some cancer cells are also thought to "differentiate" (undergo epigenetic changes) to form cancer cells with limited proliferative capacity.5,6
The implications of this cancer stem cell model for therapy are profound. For cancers that follow the cancer stem cell model, it may be necessary and sufficient to eliminate the cancer stem cells in order to cure the disease. For example, some forms of testicular cancer contain rapidly dividing undifferentiated cancer cells as well as differentiated cancer cells with little or no capacity to proliferate. In these cases, it is sufficient to eliminate the undifferentiated subset of cancer cells in order to cure the disease.7 Nonetheless, this is easier said than done as it has been proposed that cancer stem cells are more resistant to therapy than other cancer cells.8,9 Thus, it has been suggested that our inability to cure most metastatic cancers may reflect inherent therapy resistance in the cancer stem cells.
While the cancer stem cell model is an important new idea, there remain questions. Most of the best data in support of the cancer stem cell model have been from experiments in which human cancer cells were studied after transplantation into highly immunocompromised NOD/SCID mice. These mice lack T and B immune system cells, and therefore do not reject the human cells, allowing us to study the formation of human tumors in these mice. The observation from such experiments has been that only rare human cancer cells have the capacity to form tumors in such mice. As a result, it has been widely suggested that the growth and progression of many cancers are driven by rare subpopulations of cancer stem cells. However, others have raised the question of whether NOD/SCID mice underestimate the frequency of human cancer cells with proliferative potential.10,11 If so, these experiments may systematically underestimate the frequency of human cancer cells that are capable of contributing to disease progression in patients.
To test this, we undertook studies in melanoma to test whether we could identify assay modifications that would increase the percentage of human melanoma cells that could form tumors after transplantation into mice.12 When we transplanted human melanoma cells into NOD/SCID mice we found that only about 1 in a million melanoma cells were capable of forming a tumor. However, when we transplanted the same melanoma cells into a modified assay that involves the use of more highly immunocompromised mice (that lack natural killer cells in addition to T and B cells) as well as a few other changes we found that 1 in 4 cells formed a tumor. This demonstrated that tumorigenic cells are common in some human cancers, even when such cells appear rare in NOD/SCID mice. This work also demonstrated that the use of NOD/SCID mice can dramatically underestimate the frequency of human cancer cells with tumorigenic potential.
As a result of these and other studies we believe that some human cancers follow a cancer stem cell model in which disease progression is driven by small subpopulations of cancer stem cells. In contrast, we believe that many other cancers do not follow a cancer stem cell model – that tumorigenic potential is a common attribute of cells in these cancers, rather than an attribute of only a small subpopulation of cells. For cancers that fall in both categories, it will now be critical to assess the full spectrum of cells that have the potential to contribute to disease. Cancers that follow a cancer stem cell model might be treated more effectively with therapies that target cancer stem cells. In contrast, cancers in which tumorigenic potential is a common attribute of cancer cells will not be possible to treat more effectively with therapies that target rare subpopulations of cells. Thus, the question of whether a cancer has rare or common tumorigenic cells has fundamental implications for therapeutic strategies.
Given these findings, we now would like to test how frequent tumorigenic cells are in MPNSTs. Since human MPNSTs are rare and it may currently be impossible to get fresh tumor cells from multiple MPNSTs at any one medical center, we will start by addressing this issue in MPNSTs from mouse models of the disease. These studies will indicate whether many mouse MPNST cells have tumorigenic potential, or whether only a small subpopulation of MPNST cells can proliferate extensively. Our long-term plan is to attempt to confirm our results in limited numbers of human samples after results are obtained from mouse models.
We will study the MPNSTs that arise in mice that carry mutations in Nf1 and Ink4a/Arf genes or Nf1 and p53. MPNSTs in patients sometimes develop as a result of mutations that inactivate Nf1 as well as Ink4a/Arf and/or p53.13-15 We generated a mouse model of MPNST by breeding mice that have lost one copy of Nf1 to mice deficient for Ink4a/Arf and discovered that this led to the formation of MPNSTs in 25% of the adult compound mutant mice.4 Mice bearing mutations in Nf1 and p53 also develop MPNSTs during adulthood16,17 providing another mouse model to study the disease. These results demonstrate that MPNSTs develop in mice in response to similar mutations as are observed in human MPNSTs. Other mutations also play a role in this disease but it is impossible to generate and study mice with all combinations of mutations observed in patients.
We will test whether tumorigenic cells are common among freshly dissociated MPNST cells obtained from our mouse models or whether only a small, phenotypically distinct subpopulation of cells has tumorigenic potential (see Figure 3 for a description of the experimental approach). To do this we will enzymatically dissociate tumor tissue into single cells and inject limiting dilutions of tumor cells, down to a single cell, into the peripheral nerve or under the kidney capsule of mice. If a high proportion of mouse MPNST cells is tumorigenic then our results would suggest that MPNSTs are not hierarchically organized into tumorigenic and non-tumorigenic cancer cells. If many MPNST cells retain tumorigenic capacity and these cells fail to form non-tumorigenic cells this would emphasize the need to develop therapies that eliminate every MPNST cell. On the other hand, if only a small minority of MPNST cells is tumorigenic, we will attempt to prospectively identify these cells based on marker expression. That is, we will attempt to identify markers that distinguish tumorigenic from non-tumorigenic MPNST cells. Such markers could subsequently be used to facilitate efforts to develop therapies that specifically target the tumorigenic subpopulation of MPNST cells, to improve our ability to kill these cells.
Beyond the question of what fraction of MPNST cells retains tumorigenic potential, a second issue is what mechanisms regulate the proliferation of these cells. Identification of these mechanisms is a pre-requisite to develop targeted therapies to kill these cells. Since MPNSTs arise from Schwann cells, a general issue is whether the cancer cells resemble Schwann cells in the mechanisms they use to proliferate, or whether MPNST cells acquire proliferation mechanisms that resemble normal stem cells. To address this issue we will compare the gene expression profiles of mouse MPNST cells to normal neural crest stem cells, restricted Schwann cell precursors, and differentiated Schwann cells. This will enable us to globally compare the gene expression profiles of these cells as well as identifying specific pathways that may be activated in the MPNST cells. We also plan to perform functional studies on particular pathways that we hypothesize may be required to regulate the proliferation of MPNST cells. These studies will generate hypothesis for future studies of mechanisms that regulate MPNST cell survival and proliferation, and potential new therapeutic targets.
By performing these studies we will determine if mouse MPNSTs follow a cancer stem cell model and whether new therapies for MPNST should attempt to eliminate all MPNST cells or whether they should target a phenotypically distinct subpopulation of MPNST stem cells. Our experiments also have the potential to identify pathways that are activated when Schwann cells are transformed into MPNSTs, and to begin to functionally study pathways that might regulate MPNST cell proliferation. These studies may provide new mechanistic insights into the development of MPNSTs as well as generate hypotheses for future studies into novel mechanisms/therapeutic targets.
Thanks to the Liddy Shriver Sarcoma Initiative for supporting this work.
By Johanna Buchstaller, PhD
and Sean J. Morrison, PhD
Howard Hughes Medical Institute
Department of Internal Medicine
Life Sciences Institute
University of Michigan
Ann Arbor, Michigan
1. Rubin JB, Gutmann DH. (2005). Neurofibromatosis type 1 - a model for nervous system tumour formation? Nat Rev Cancer 5:557-564.
2. Williams VC, Lucas J, Babcock MA, Gutmann DH, Korf B, Maria BL. (2009). Neurofibromatosis type 1 revisited. Pediatrics 123:124-133.
3. Zheng H, Chang L, Patel N, Yang J, Lowe L, Burns DK, Zhu Y. (2008). Induction of abnormal proliferation by non-myelinating Schwann cells triggers neurofibroma formation. Cancer Cell 13:117-128.
4. Joseph NM, Mosher JT, Buchstaller J, Snider P, McKeever PE, Lim M, Conway SJ, Parada LF, Zhu Y, Morrison SJ. (2008). The loss of Nf1 transiently promotes self-renewal but not tumorigenesis by neural crest stem cells. Cancer Cell 13:129-140.
5. Pardal R, Clarke MF, Morrison SJ. (2003). Applying the principles of stem-cell biology to cancer. Nature Cancer Reviews 3:895-902.
6. Reya T, Morrison SJ, Clarke MF, Weissman IL. (2001). Stem cells, cancer, and cancer stem cells. Nature 414:105-111.
7. Horwich A, Shipley J, Huddart R. (2006). Testicular germ-cell cancer. Lancet 367:754-765.
8. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, Hjelmeland AB, Dewhirst MW, Bigner DD, Rich JN. (2006). Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444:756-760.
9. Diehn M, Cho RW, Lobo NA, Kalisky T, Dorie MJ, Kulp AN, Qian D, Lam JS, Ailles LE, Wong M, et al. (2009). Association of reactive oxygen species levels and radioresistance in cancer stem cells. Nature.
10. Kelly PN, Dakic A, Adams JM, Nutt SL, Strasser A. (2007). Tumor growth need not be driven by rare cancer stem cells. Science 317:337.
11. Williams RT, den Besten W, Sherr CJ. (2007). Cytokine-dependent imatinib resistance in mouse BCR-ABL+, Arf-null lymphoblastic leukemia. Genes Dev 21:2283-2287.
12. Quintana E, Shackleton M, Sabel MS, Fullen DR, Johnson TM, Morrison SJ. (2008). Efficient tumour formation by single human melanoma cells. Nature 456:593-598.
13. Agesen TH, Florenes VA, Molenaar WM, Lind GE, Berner JM, Plaat BE, Komdeur R, Myklebost O, van den Berg E, Lothe RA. (2005). Expression patterns of cell cycle components in sporadic and neurofibromatosis type 1-related malignant peripheral nerve sheath tumors. J Neuropathol Exp Neurol 64:74-81.
14. Menon AG, Anderson KM, Riccardi VM, Chung RY, Whaley JM, Yandell DW, Farmer GE, Freiman RN, Lee JK, Li FP, et al. (1990). Chromosome 17p deletions and p53 gene mutations associated with the formation of malignant neurofibrosarcomas in von Recklinghausen neurofibromatosis. Proc Natl Acad Sci U S A 87:5435-5439.
15. Perrone F, Tabano S, Colombo F, Dagrada G, Birindelli S, Gronchi A, Colecchia M, Pierotti MA, Pilotti S. (2003). p15INK4b, p14ARF, and p16INK4a inactivation in sporadic and neurofibromatosis type 1-related malignant peripheral nerve sheath tumors. Clin Cancer Res 9:4132-4138.
16. Cichowski K, Shih TS, Schmitt E, Santiago S, Reilly K, McLaughlin ME, Bronson RT, Jacks T. (1999). Mouse models of tumor development in neurofibromatosis type 1. Science 286:2172-2176.
17. Vogel KS, Klesse LJ, Velasco-Miguel S, Meyers K, Rushing EJ, Parada LF. (1999). Mouse tumor model for neurofibromatosis type 1. Science 286:2176-2179.
V6N1 ESUN Copyright © 2009 Liddy Shriver Sarcoma Initiative.
The Liddy Shriver Sarcoma Initiative funded this $50,000 grant in February 2009. The study was made possible, in part, by generous donations made in loving memory of James Stulce, who lost his life to this disease.
An extensive report of this study's results was published in the journal Cancer Cell in February 2012.
A. During embryonic development, neural crest stem cells leave the neural tube and migrate into peripheral nerves, where they differentiate into Schwann cells. B. In adult nerves, myelinating Schwann cells (M) wrap around large caliber axons (A) to form the myelin sheath that insulates nerves and promotes the transmission of nerve impulses (note that Figure B is an electron micgrograph with 5800x magnification).
Figure 2A (left). Stochastic/clonal evolution model: Cancer cells are heterogeneous, but most cells can proliferate extensively and form new tumors. All cancer cells must be eliminated to cure the patient. Figure 2B (right). Cancer stem cell model: Cancer cells are heterogeneous and only the cancer stem cells (csc; green) have the ability to proliferate and form new tumors. Other cancer cells have little capacity to divide. To cure cancer it may be necessary and sufficient to eliminate the cancer stem cells.
A single cells suspension is prepared from fresh tumor tissue, and distinct subpopulations of cancer cells that differ in terms of marker expression (M) are transplanted into mice. If one cell fraction preferentially gives rise to tumors, we conclude that the cancer follows a cancer stem cell model. If all fractions give rise to tumors, we conclude that the cancer follows a stochastic model and that many cells have the capacity to proliferate extensively.
Figure 1A (left): Cross section through mouse. MPNSTs stained with hematoxylin/eosin to visualize cell bodies (pink) and nuclei (purple). The arrow marks a cell undergoing mitosis. Figure 1B (right): Cell nuclei are stained in brown. Arrows mark proliferating cells. Magnification is 40x.
Figure 2A&B: Mouse MPNSTs are enzymatically dissociated into single cells. Figure 2C: Single MPNST cells are seeded into culture medium and cell colonies injected into the kidney capsule of mice. Figure 2D: Alternatively, small numbers of freshy isolated MPNST cells are directly transfered into mice without prior culture. The mice are monitored for tumor formation (white arrows) for five months.